Paraboloid antenna surfaces suffer performance degradation due to structural deformation. A first step in the prediction of the performance degradation is to find the best-fit paraboloid to the deformed surface. Examined here is the question of whether rigid body translations perpendicular to the axis of the paraboloid should be included in the search for the best-fit paraboloid. It is shown that if these translations are included the problem is ill-conditioned, and small structural deformation can result in large translations of the best-fit paraboloid with respect to the original surface. The magnitude of these translations then requires nonlinear analysis for finding the best-fit paraboloid. On the other hand, if these translations are excluded, or if they are limited in magnitude, the errors with respect to the restricted not-so-best-fit paraboloid can be much greater than the errors with respect to the true best-fit paraboloid.